Investigation of a Quasi-Integral Sliding Mode Control for a nonlinear Maglev experimental system

This paper investigates a robust control technique for a magnetic levitation (Maglev) system designed for practical use, addressing common failures of payload-specific systems that operate only under ideal lab conditions. A payload-agnostic maglev system is developed to handle external perturbations and noise. A high-fidelity non-linear electromechanical-coupled model of a ferromagnetic beam and electromagnets is developed by empirically determining key model parameters. Due to the intrinsic instability and nonlinearity of maglev systems, high sampling frequencies are necessary for effective stabilization, making complex controllers impractical. A Quasi-Integral Sliding Mode Controller (QISMC) is proposed, balancing simplicity and robustness while addressing chattering and steady-state error issues typical of conventional Sliding Mode Controllers (SMC). The QISMC's performance is compared to a linear PIDN controller, which, despite its effectiveness in limited ranges, underperforms in dynamic environments whereas the QISMC excels. The control algorithms are implemented on a dedicated embedded system using Simulink Desktop Real-Time software, and experimental results confirm the simulations' accuracy, showing strong consistency between theoretical predictions and real-world performance.